View Issue Details

IDProjectCategoryView StatusLast Update
0000267OpenFOAMBugpublic2011-09-08 12:43
Reporteruser171Assigned Tohenry  
PrioritynormalSeverityminorReproducibilityalways
Status resolvedResolutionfixed 
Platformx84_64OSUbuntuOS Version11.04
Summary0000267: zeta equation and fMu function in qZeta turbulence model wrong
DescriptionThe implementation of the zeta equation is wrong in my opinion. It can be only a typo or by intention, but in the zeta equation

tmp<fvScalarMatrix> zetaEqn
    (
        fvm::ddt(zeta_)
      + fvm::div(phi_, zeta_)
      - fvm::laplacian(DzetaEff(), zeta_)
     ==
        (2.0*C1_ - 1)*G*zeta_/q_
      - fvm::Sp((2.0*C2_ - dimensionedScalar(1.0))*f2()*zeta_/q_, zeta_)
      + E
    );

the second to last term

- fvm::Sp((2.0*C2_ - dimensionedScalar(1.0))*f2()*zeta_/q_, zeta_)

should be

- fvm::Sp((2.0*C2_*f2()_ - dimensionedScalar(1.0))*zeta_/q_, zeta_)

in order to reproduce the relation

C_zeta f_zeta = 2 C_eps f_eps - 1 as described in the paper just above equation (16)

The other thing is an extra turbulence Reynolds number dependent factor in the model, that appears in the the fMu() function

           exp(-6.0/sqr(scalar(1) + Rt/50.0))
           *(scalar(1) + 3.0*exp(-Rt/10.0));

The second factor scalar(1) + 3.0*exp(-Rt/10.0)) does not exist in the relation given in equation (18) of the paper.
Additional InformationPaper referred to is
Gibson and Dafa'Alla, Two-Equation Model for Turbulent Wall Flow, AIAA Journal, vol. 33, No. 8, 1994
TagsModelling, turbulence models

Activities

henry

2011-09-08 12:43

manager   ~0000645

The implementation of the q-zeta in OpenFOAM is based on the final technical report on the work by Dafa'Alla, Juntasaro and Gibson which was subsequently presented and published:

        "Calculation of oscillating boundary layers with the
        q-[zeta] turbulence model"
        A.A. Dafa'Alla, E. Juntasaro and M.M. Gibson

        Engineering Turbulence Modelling and Experiments 3:
        Proceedings of the Third International Symposium,
        Crete, Greece, May 27-29, 1996, p. 141.

        Editors: Wolfgang Rodi and G. Bergeles

in which alternative forms of the damping functions are presented.
 
However
- fvm::Sp((2.0*C2_ - dimensionedScalar(1.0))*f2()*zeta_/q_, zeta_)

should be

- fvm::SuSp((2.0*C2_*f2() - dimensionedScalar(1.0))*zeta_/q_, zeta_)

and this is resolved by commit ebe324f10a942fd92f0fbb0aa9dade8765a18a90

Issue History

Date Modified Username Field Change
2011-08-03 11:57 user171 New Issue
2011-08-03 11:59 user171 Tag Attached: Modelling
2011-08-03 11:59 user171 Tag Attached: turbulence models
2011-09-08 12:43 henry Note Added: 0000645
2011-09-08 12:43 henry Status new => resolved
2011-09-08 12:43 henry Resolution open => fixed
2011-09-08 12:43 henry Assigned To => henry